Damage localization in plate-like structures using time-varying feature and one-dimensional convolutional neural network

•A new time-varying damage index that preserves temporal information was proposed.•A convolutional neural network was developed to further extract temporal features.•Damage localization was realized using four transducers and limited training data.•Good transferability and robustness of the CNN were demonstrated.•Once constructed, the CNN can predict damage location in milliseconds.